Cloud computing is a new computing
model in which the services are rented to the customers on demand. We connect
to many websites via internet without having to know about the infrastructure
details. In cloud computing applications are offered as services and it can be
used via internet. Not only desktops, laptops but also mobile phones can be
connected. For companies, using cloud computing will reduce the cost. There
will be service level agreements for the services provided. Amazon, IBM,
Google, Microsoft all own their data centers and host applications. Cloud
computing refers to both services offered via internet and hardware and
software in datacenters.
are not required to build and maintain complex infrastructure. They can pay as
they use the cloud computing services. Resources can be allocated and
de-allocated on demand. Hence resources are not always required as when they
are required for a peak load. This lowers the operating cost. Service provider
can scale up according to the demands. Consumers are now having infinite
capacity. There are three types in cloud computing Infrastructure as a service,
Software as a service and Platform as a service.
Elastic Compute Cloud (EC2) looks like hardware. It will provide virtual
computing environments called instance. Users need not invest in hardware.
Users can upload software to the instance and can modify it. Instances can be
placed in multiple locations. Google AppEngine is used to develop and host web
applications. Windows Azure is platform as a service cloud provider.
discuss the research challenges of cloud computing.
Automated service provisioning: – The provider must
adjust based on the requests and meet the SLA. It must develop a predictive
tool which can predict the future demands and allocate the resources according
to the demand.
Virtual Machine Migration: – It requires
extremely short downtimes. However, this is not straightforward.
Server consolidation: – Using this approach we
can maximize the resource utilization and minimize energy consumption.
Energy Management: – we have to trade-off
between savings and performance
Data Security: – Service provider must depend on
infrastructure provider to secure the data
Software Frameworks: – MapReduce frameworks
consume significant CPU resources
Software technologies and data management: – The
file systems used by Hadoop, Dryad are different from traditional file systems
Novel cloud architectures: – Small data centers
are advantageous compared to big data centers
Availability / Business continuity
Data Lock in
is a promising field in IT. There are many key challenges that needs to resolved
by research community.